What to count (was: Hair-splitting...)

Daniel Riaño danielrr at mad.servicom.es
Wed Jul 14 07:27:21 EDT 1999

(sorry about the delay --busy week here)

Paul S. Dixon wrote:
>If the perfect and pluperfect are rare in the Koine, and if the
>"aorist has effectively assumed their functions in narrative,"
>then one would certainly expect that such rarity would be
>evidenced by a comparison of the stats, regardless of the
>nuances of the perfect (which I don't think have any bearing
>upon the question).  If not, then what is the source of
>such suspicions?

	There is nothing simple in counting grammatical units. If you want
your data to be of some scientific value, you have to use (and explain to
the readers) a number of conventions, usually quite a bunch. Once you have
applied systematically a sound method in a text corpus long enough, you
still must face to the question when the discrepancies in the use of some
linguistic device between two authors is relevant or not: at least for this
last question you can apply the chi square text, to see if such
discrepancies are due to happenance.
	Now to your question: I think that if you compare statistics about
the use of a given morphological category in two authors (or two sets of
authors) as different as Homer and Thucydides, or Thucydides an NT
*without* having in account further analysis (such as nuances of meaning,
and other: see infra) your results will be most probably meaningless. I
think I already wrote to the list that the use of the dative in Luke is far
larger than the use of the same case in Polybius (in absolute numbers,
using as a sample of the same extension). Does it means that the dative was
more productive in Luke, or that the Dative case experienced an improbable
anastasis in Luke's time? No way: you need to further study the kind of
words that are put in the dative, and the set of functions they serve (very
limited in Luke in both cases), plus the semantic range of the regent, etc,
to start understanding what the raw number means.

Carl W. Conrad Wrote:
I don't think that the ratios  of aorists to perfects and pluperfects would
be that significant in themselves; I think analysis is required.
(and in the same line) Edgar M. Krenz wrote:
>This is a very interesting thread. I wonder, however, whether some of these
>statistics might not be deceptive.

	I think so. Y. Duhoux, the author of the book I quoted is
undoubtedly one of the authors who has studied more profoundly (and
extensively) aspect in the Greek verb (but preferably Classical and Homeric
Greek). The statistics he borrowed for the 1992 book where a bit
unpolished, but later he has published of very interesting papers in the
subject where he uses his own account of data, done with greater care. In a
paper I recommend (Duhoux, Y. "Études sur l'aspect verbal en grec ancien,
1: presentation d'une methode." BSL 90 [1995]: 241-299) he wrote what can
be consider a "methodological manifesto" for this kind of studies. Inter
alia, and about this question he writes:
	"Dans une étude aspectuelle, il importe de ne faire porter l'examen
que sur les verbes qui autorisent un choix aspectuel complet, c'est-à-dire
la possibilité théorique d'utilliser les thèmes de présent, d' aoriste et
de parfait. Autrement, on court un risque certain de biaiser les données"
(p. 243). (The unquoted part of C. W. Conrad posting of Sat, 10 Jul 1999
17:04:32 -0400 is in the same line of resoning)

	For his research (see also: Duhoux, Y. "L'aspect verbal en Grec
ancien: premieres leçons d'une recherche." SYNTANTIKA .7 (1994): 9-18.)
Prof. Duhoux was using a database where, every apparition of a verb form
was introduced in a new record with about 70 parameters (p. 242), and I
know that his elevées who are writing their dissertations about verbal
aspect have enlarged the list of parameters.
	A conclusion: the study of any grammatical category requires a very
minutious description of the syntactical behaviour and context in every
instance. My own conclusion is that even a so rigorous description is
insufficient if you are using an ordinary relational database (like
FileMaker, FoxPro, etc), which necessary leaves out hard-to-annotate
parameters relative to word order, etc, (and that's why I had to build my
own syntactical editor and data base for the study of case.)

> I think that it should be restricted to
>the indicative mood. For example, one would get a preponderance of aorist
>infinitive forms of  LAMBANW over present and perfect infinitives, simply
>because that is the infinitive form of preference.
>At the lease, Daniel, I would be interested in knowing the statistics for
>the different moods. The rarity of any perfect infinitives would skew the
>interpretation, IMHO. ;-)

	Right. In reality, the stats I gave were *the numbers corresponding
only to the indicative case*. Sorry again: I forgot to mention that. But
for you to compare, now I am giving the absolute numbers, i.e. regardless
of mood. Same work, same authors, same pages. Caution: use at your own risk

			(all moods)
		Aor.		Perf./Pl.Perf.

Ilias		9380	85.21%	1628	20.48%
Odyssea		7346	85.42%	1254	18.42%
Herodotus		11885	83.02%	2430	13.20%
Thucydides	8398	86.25%	1339	15.11%
Att. Aut.		1396	79.00%	371	12.87%

		Aor.Ind.		Perf./Pl.Perf. Ind.

Ilias		5334	86.27%	849	13.73%
Odyssea		3944	84.89%	702	15.11%
Herodotus		4900	81.31%	1126	18.69%
Thucydides	4054	89.12%	495	10.88%
Att. Aut.		398	70.19%	169	29.81%

Daniel Riaño Rufilanchas
Madrid, España

More information about the B-Greek mailing list